Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A practical, actionable reference with executable JSON and SQL recipes and strong guardrails in the Gotchas section. It is mostly concise and well-structured, though the orientation prose could be tighter and the long inline tables suggest a possible split for a future reference file.
Suggestions
Tighten the intro paragraph — drop 'This skill teaches you (the agent) how to use that classification to:' and lead directly with the capability list to save tokens.
Consider moving the full $virt_* property table and HogQL function table into a references/ file, leaving SKILL.md as a concise overview with the key examples inline.
Add a short 'verify the result' note to one recipe (e.g. confirm human pageviews drop noticeably after applying the $virt_is_bot filter) to add an explicit checkpoint.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean and free of padding about concepts Claude already knows, but the orientation paragraph ('This skill teaches you (the agent) how to use that classification to:') and a few explanatory asides could be trimmed. | 4 / 5 |
Actionability | Provides copy-paste-ready JSON property filters and TrendsQuery examples plus complete SQL with the coalesce(nullIf(...)) user-agent expression, covering the common cases (exclude, measure share, list bots, AI-agent slice, raw SQL). | 5 / 5 |
Workflow Clarity | Recipes are clear single actions with a useful Gotchas section acting as guardrails (e.g. 'Don't silently drop the host filter'), but there are no explicit validate-then-retry checkpoints; none are strictly required since the queries are read-only. | 4 / 5 |
Progressive Disclosure | Well-organized into clear sections (When to use, classification surface, traffic types, recipes, gotchas) with no bundle files to delegate to; the full property/function tables are reasonable inline but the body is longer than a minimal overview. | 4 / 5 |
Total | 17 / 20 Passed |